Staff AI Systems Engineer - RL, Infrastructure & Production

Ambral Labs; San Francisco (CA); On-site; USD 150,000 - 230,000; Full time

Benefits offered by this job

Equity ownership

Equinox membership

Free meals, coffee, snacks

Health insurance

Unlimited PTO

Job summary

Ambral Labs is building replayable environments and eval sets grounded in real workflows to improve model performance for enterprise tasks. The team owns the core infrastructure to test policies and agents against authentic historical data in production-like settings.

You will collaborate across research, infrastructure, and production systems, deploying into real enterprise workflows and measuring outcomes with rigorous observability, reproducibility, and cost-aware design.

Qualifications

• 1-7 years of experience building production software or machine-learning systems.
• Bonus points for RL environments, LLM post-training, evaluation infrastructure, agent harnesses.
• Understand how environment design, reward design, context, tooling, and policy behavior interact.
• Able to turn fuzzy business objectives into tasks and signals that can be evaluated reliably.
• Able to diagnose whether a model’s limitations come from the model itself, its context, its tools, its harness, or its training.

Responsibilities

• Building an environment factory that converts recorded enterprise data and task definitions into runnable environments.
• Designing graders that turn ambiguous business objectives into verifiable rewards.
• Developing methods for mining useful tasks, trajectories, and evaluation cases from historical workflows.
• Creating eval sets that are representative, reproducible, and resistant to overfitting.
• Finding the right combinations of models, tools, context, and policies to maximize performance while reducing inference cost.
• Training and evaluating agents that operate over long horizons, incomplete information, and large tool spaces.
• Building replay and observability systems that make agent behavior explainable and measurable.
• Scaling from individual environments to thousands of concurrent training and evaluation runs.

Skills

Production software

Machine-learning systems

Reinforcement learning

Evaluation infrastructure

Agent harnesses

System scalability

Observability

Reproducibility

Job description

Ambral Labs is building replayable environments and eval sets grounded in real workflows to improve model performance for enterprise tasks. The team owns the core infrastructure to test policies and agents against authentic historical data in production-like settings.

You will collaborate across research, infrastructure, and production systems, deploying into real enterprise workflows and measuring outcomes with rigorous observability, reproducibility, and cost-aware design.